MIT Open Access Articles

Analysis of reversible ejectors and definition of an ejector efficiency

The MIT Faculty has made this article openly available. Please share
how this access benefits you. Your story matters.

Citation: McGovern, Ronan K., G. Prakash Narayan, and John H. Lienhard. “Analysis of
Reversible Ejectors and Definition of an Ejector Efficiency.” International Journal of Thermal
Sciences 54 (April 2012): 153–166.

As Published: http://dx.doi.org/10.1016/j.ijthermalsci.2011.11.003

Publisher: Elsevier

Persistent URL: http://hdl.handle.net/1721.1/102358

Version: Author's final manuscript: final author's manuscript post peer review, without
publisher's formatting or copy editing

Terms of use: Creative Commons Attribution-NonCommercial-NoDerivs License

 Analysis of Reversible Ejectors and Definition of an Ejector
Efficiency
Ronan K. McGovern, G. Prakash Narayan, John H. Lienhard V∗
Center for Clean Water and Clean Energy, Department of Mechanical Engineering, Massachusetts Institute
of Technology, Cambridge, MA 02139, USA.

Abstract
Second Law analyses of ejector performance have rarely been conducted in literature.
Many measures of ejector efficiency have not always been clearly defined and the rationale
underlying and justifying current performance metrics is often unclear. One common means
of assessing performance is to define a thermodynamically reversible reference process against
which real processes may be benchmarked. These reversible processes represent the thermo-
dynamic limit of real ejector performance. In this paper parameters from real and reversible
processes are compared and performance metrics are defined. In particular, the entrainment
ratio of real devices is compared to the reversible entrainment ratio and denoted the reversible
entrainment ratio efficiency. An efficiency comparing the ejector performance to that of a
turbine-compressor system is also investigated, as is an exergetic efficiency. A rigorous anal-
ysis of performance metrics reported in the literature is undertaken. Graphical illustrations
are provided to support intuitive understanding of these metrics. Analytical equations are
also formulated for ideal-gas models. The performance metrics are then applied to exist-
ing experimental data to illustrate the difference in their numerical values. The reversible
entrainment ratio efficiency ηRER is shown to always be lower than the turbine-compressor
efficiency ηTER . For general air-air and steam-steam ejectors, the exergetic efficiency ηX is
very close in numerical value to the reversible entrainment ratio efficiency, ηRER .
Keywords: Ejector, Thermocompressor, Pressure exchange, Second Law of
Thermodynamics, Isentropic efficiency

R.K. McGovern, G.P. Narayan, and J.H. Lienhard V, “Analysis of Reversible Ejectors and Definition
of Ejector Efficiency,” International Journal of Thermal Sciences, 54(4):153–166, April 2012.

∗
[email protected]

Preprint submitted to International Journal of Thermal Sciences April 11, 2016

Nomenclature
Symbols
cp specific heat capacity at constant pressure [kJ/kg·K]
ER entrainment ratio [-]
h specific enthalpy [kJ/kg]
h̃ specific molar enthalpy [kJ/kmol]
hE,D specific enthalpy of the entrained fluid at the discharge [kJ/kg]
hM,D specific enthalpy of the motive fluid at the discharge [kJ/kg]
j dissipation [J]
ṁ mass flow rate [kg/s]
M molar mass [kg/kmol]
Ṅ molar flow rate [kmol/s]
p partial pressure [bar or kPa]
P absolute pressure [bar or kPa]
Q̇ rate of heat transfer [W]
R gas constant [kJ/kg·K]
RER Reversible Entrainment Ratio [-]
s specific entropy [kJ/kg·K]
s̃ specific molar entropy [kJ/kmol·K]
Ṡgen entropy generation rate [kW/K]
s̄gen dimensionless entropy generation [-]
T temperature [ ◦ C or K]
v specific volume [m3 /kg]
x dryness fraction of steam [-] or specific exergy [J/kg]
y useful work done [J]

Greek
γ ratio of specific heats
∆ change
η efficiency [-]
Πc compression ratio (PD /PE ) [-]
φ relative humidity [-]
ω absolute humidity [kg H2 O/kg carrier gas]
ξ ratio of moles of vapor to moles carrier gas [-]

Subscripts
c compression
cg carrier gas
comp compressor
D discharge fluid
E entrained fluid
f saturated liquid
g saturated vapour
irrev associated with irreversible dissipation

2
M motive fluid
NE nozzle exit
q associated with heat transfer
RDP reversible discharge pressure
RER reversible entrainment ratio
RHE reversible heat engine
s isentropic
TER turbine-compressor entrainment ratio
turb turbine
vap vapor
X exergetic
0 ambient

Superscripts
rev reversible
sat saturated

3
1. Introduction
Steady flow ejectors are devices without moving parts that typically combine two fluids,
one of high total pressure and one of low total pressure, to obtain a fluid at an intermediate
total pressure at the discharge. The fluid with high total pressure, known as the motive fluid,
is expanded through a nozzle to high velocity and low pressure (point NE in Fig.1). The low
pressure at the nozzle exit provides the driving force for suction of the low pressure fluid,
known as the entrained fluid, at pressure PE , such that PM > PE > PNE . The motive and
entrained streams undergo mixing, brought about by shear forces between the fluids, before
the resulting mixture is diffused to attain the discharge pressure, PD , Fig. 1.

[Figure 1 about here.]

The most intuitive although approximative manner to comprehend ejector operation is

to analyse performance using simple 1-D analyses. Examples of excellent work in this regard
have been done by Elrod [1], Keenan et al. [2], Chou et al. [3] and Arbel et al. [4]. Ejectors
are relatively simple in design and can provide reliable operation at low capital cost. These
advantages have led to ejectors being used in processes requiring the drawing of a vacuum
(e.g. extraction of non-condensible gases from condensers) or the compression of a gas (e.g.
within refrigeration cycles). In thermal desalination, ejectors are commonly employed to
compress and entrain water vapor. Such devices are known as thermal vapor compressors.
The increased use of thermal vapor compressors for energy intensive processes of thermal
desalination, in particular multi-effect distillation [5–8], has brought about renewed interest
in the efficiency of ejectors. Despite the interest in energy efficiency, scant attention has been
paid to the definition and justification of the utility of ejector performance metrics employed
to date. Bulusu et al. [9] presented common measures of ejector performance, whilst Arbel
et al. [4] performed an analysis of entropy generation within ejectors. Nonetheless, there
remains no universally accepted efficiency to characterise and compare the performance of
such devices.
In this paper, ejector performance is evaluated by benchmarking real ejector processes
against carefully defined thermodynamically reversible processes. In Section 2, a systematic
methodology is presented to define such processes. One important parameter arising from
this analysis is the reversible entrainment ratio, i.e. the maximum theoretical entrainment
ratio achievable by a thermodynamically reversible ejector. The variation in the reversible
entrainment ratio with an ejector’s operating conditions is investigated in Section 3. Then,
in Section 4 ejector performance metrics are explained and justified by means of analytical
equations or graphical methods. Finally, in Section 5, the numerical values of these metrics
are computed for experimental data available in open literature. The exergetic efficiency ηX
of an ideal gas ejector with inlet fluids at the same temperature is shown to be identical to
the reversible entrainment ratio efficiency ηRER .

2. Description of Reversible Ejector Processes

This section establishes a basis for benchmarking real device performance against the
performance of ideal, thermodynamically reversible processes. In order to effectuate this
comparison, the following steps are necessary:

4
 1. Identify the input quantities, output quantities and equations describing a thermody-
namically reversible process.
2. Define a thermodynamically reversible reference process, against which real processes
may be compared.
3. Develop a performance metric based on a comparison of parameters between the real
and reversible processes. (4)

A control volume model of an ejector process is presented in Fig. 2. The inlet stream
of higher pressure is known as the motive flow, the low pressure inlet stream is termed the
entrained flow and the outlet stream is termed the discharge flow. The control volume in
both real and reversible cases is considered externally adiabatic (Q̇0 = 0). Due to the high
flow speeds within ejectors, the rate of heat loss from the device to ambient is taken to be
small relative to changes of enthalpy of the streams within. The nomenclature of Fig. 2 shall
henceforth be used in refering to the mass flow rates and state properties of inlet and outlet
streams. All analyses shall consider only equilibrium states of inlet and outlet fluids where
the fluids are at conditions of stagnation (zero kinetic energy). Thus, all temperatures and
pressures are total rather than static temperature and pressures.

[Figure 2 about here.]

2.1. Equations Describing a Thermodynamically Reversible Process

Taking a control volume approach, the First and Second Laws of Thermodynamics may
be written for a real ejector process involving pure and identical fluids at the ejector inlet:
First Law of Thermodynamics for an adiabatic process:

ṁM hM + ṁE hE = ṁD hD (1)

hM + ER · hE = (1 + ER)hD (2)
The entrainment ratio, ER is defined as the ratio of entrained mass flow to motive mass flow,
ṁE
ER = (3)
ṁM
Second Law of Thermodynamics for an adiabatic process:

ṁM sM + ṁE sE + Ṡgen = ṁD sD (4)

Ṡgen
sM + ER · sE + = (1 + ER)sD (5)
ṁM
Noting that no entropy is generated during a reversible process, the equations for an adiabatic
and reversible process may be formulated as follows:

hM + RER · hE = (1 + RER)hrev
D (6)

5
 sM + RER · sE = (1 + RER)srev
D (7)
Reversible entrainment ratio RER is defined as the ratio of entrained mass flow to motive
mass flow in the reversible process:
 rev
ṁE
RER = (8)
ṁM

For a pure fluid, the thermodynamic state is fully specified at the control volume inlets
and outlet once two state properties (i.e. h and s) are known. An equation of state may be
written relating the temperature T or the pressure P to the specific enthalpy and entropy.
The specific enthalpy and entropy of the discharge stream is simply the mass weighted average
of the inlet streams, see Fig. 3. Of course, according to Eq. (6) and (7), the entrainment ratio,
ER, is the mass ratio relevant to this weighting. Hence, on a diagram of specific enthalpy
versus specific entropy, the discharge fluid state must lie on the line joining the states of the
inlet fluids, as shown in Fig. 3. The next step is to define the different possible reversible
reference processes.

[Figure 3 about here.]

2.2. Definition of Thermodynamically Reversible Reference Processes

Six state variables and one mass flow rate ratio describe the process of Fig. 2. For a
real process there are two equations, Eq. (2) and Eq. (5), including one further unknown,
Ṡgen
ṁM
. This value of this entropy generation term is dictated by the quality of design of the
device within the control volume. For a reversible process, there are two associated equations,
Eq. (6) and Eq. (7). Consequently, given 7 variables and 2 equations, there are five degrees
of freedom that must be specified in order to define a reversible reference process. There
are therefore 21 (7-choose-5) different ways in which a reference process could be defined. A
select number of combinations are discussed in the following subsections.

2.2.1. Reversible Discharge Pressure Reference Process

In a paper elucidating the sources of irreversibilty within steady flow ejectors, Arbel
et al. [4] investigated ejector performance in great detail by comparing real processes to a
reversible reference process, described in Table 3:

[Table 1 about here.]

In essence, Arbel et al.’s reference processes asks: for fixed inlet fluid states and a fixed
entrainment ratio, what is the maximum discharge pressure achievable if the process is re-
versible? On an h − s diagram, Arbel’s process may be represented as:

[Figure 4 about here.]

Discharge enthalpy in both the real and reference case is a weighted average of the inlet
fluid enthalpies. Since by definition of this reference process, the real and reversible entrain-
ment ratios are equal, the discharge specific enthalpy is also equal in the real and reversible

6
cases, Eq. (2). In the real process, entropy is generated and thus the discharge specific
entropy is greater than the weighted mean of the entropy of the inlet streams.
For applications where the inlet states are fixed, as is the relative mass flow of motive
and entrained fluid available, this reference process may be employed to determine how close
the actual discharge pressure achieved is to the reversible limit.

2.2.2. Reversible Entrainment Ratio Reference Process

In many situations, the inlet states to the ejector are known, as is a desired discharge
pressure1 . Such is the case when one has steam available at a fixed state and wishes to
compress vapor, also at a known state, to a desired pressure. The efficiency of the ejector
will determine the mass flow of vapor one can entrain per unit mass flow of steam (i.e.
the entrainment ratio). In such a scenario, the reference process described in Table 4 is of
relevance:

[Table 2 about here.]

On an h − s diagram, the reversible entrainment ratio reference process may be represented

as:

[Figure 5 about here.]

The real discharge pressure in Fig. 5 is the same as that in Fig. 4. However, the discharge
state is different. In the reversible entrainment ratio reference process, the reversible and
real discharge pressure are set as equal. Employing the reversible entrainment ratio reference
process, one may compare the real entrainment ratio to the entrainment ratio achieved in a
reversible process.

2.2.3. Summary and Further Reference Processes

The above two reference processes are only two of the 21 possible reference processes for
a dual inlet, pure and identical fluid ejector. In particular, knowledge of the entrainment
ratio is interesting from a cost perspective as it quantifies the mass of fluid entrained per
unit mass of motive fluid supplied. The reversible discharge pressure reference process is also
considered due to the notable work presented on this process by Arbel et al. [4]. One could
also consider a reference process whereby the motive and discharge state are fixed along with
the entrainment ratio, and the real entrained pressure and temperature are compared to the
entrained pressure and temperature obtained in a reversible process. Having established a
basis for the definition of such processes, further permutations are left to the reader bearing
in mind that the choice of reference process must be guided by the intended application of
the ejector.

1
Where one is dealing with an ejector involving fluids of identical chemical composition, it is equivalent to
define either the discharge pressure or the saturation temperature at that pressure, as is done by Al Khalidy
and Zayonia [10] and by Elrod [1].

7
2.3. Interpretation of a Thermodynamically Reversible Ejector Process
Here a thermodynamically reversible ejector process is further elucidated. Importantly,
this illustration does not attempt to explain how a reversible process could be achieved
in a real steady-flow ejector. Rather, it points out how, theoretically, a reversible process
could take place within the control volume of an ejector. For simplicity, let us consider an
ejector process with pure fluids at the inlet, both of identical chemical composition. The
a reversible ejector process can now be interpreted as the sum of two processes. The first
process establishes mechanical equilibrium and the second establishes thermal equilibrium.
The process that establishes mechanical equilibrium at the discharge pressure, PD is in fact
a turbine-compressor processs. The motive and entrained fluids are respectively expanded
and compressed through an adiabatic and isentropic turbine and compressor, Fig. 6.

[Figure 6 about here.]

The thermal equilibrating process may be represented by a reversible heat engine that trans-
fers heat (and entropy) from the the hotter of the fluids M’ or E’ to the colder fluid, allowing
them to reach an equilibrium temperature TD . Of course, work is produced by the heat engine
to supply the compression process. It is therefore clear, that in a reversible ejector process,
the work available for compression is greater than that available in a turbine-compressor
process. Consequently, the reversible entrainment ratio will always be greater than or equal
to the turbine-compressor entrainment ratio. In the case where the fluid exiting the turbine
and the compressor in Fig. 6 are at the same temperature, no work will be done by the re-
versible heat engine. Consequently, in this special case, the reversible entrainment ratio and
the turbine-compressor entrainment ratio will be identical. The equilbrium process paths of
Fig. 6 are represented in Fig. 7.

[Figure 7 about here.]

Until this stage, processes involving only pure and identical inlet fluids have been con-
sidered. However, the same analysis can be generalised to cover processes such as the com-
pression of a moist air stream using high pressure steam. In such cases, additional variables,
such as humidity ratio, are required to fully define the equilibrium state at any point.

3. Reversible Entrainment Ratio Calculations for Different Fluids

At present, the performance of ejectors is typically predicted using graphical methods
or by semi-empirical correlations [11, 12]. These methods allow the calculation of the en-
trainment ratio as a function of inlet and outlet fluids’ thermodynamic states. Whilst the
reversible entrainment ratio does not provide much information about the entrainment ratio
of real devices, it does provide an upper bound upon the entrainment ratio achievable. In this
section, trends in the reversible entrainment ratio with inlet fluid conditions and discharge
pressure are illustrated for ideal gases (air-air), steam-steam and steam-moist-carrier-gases
ejectors.

8
3.1. Air-air
3.1.1. Derivation of the Reversible Entrainment Ratio for Ideal Gas Ejectors
For two reasons, the ideal gas ejector model is central to the understanding of ejector
operation and performance characterization. First, the ideal gas assumptions allow analytical
expressions to be derived for ejector performance. Secondly, for performance characterization
of ejectors, the ideal gas model is appropriate for many ejector processes, as will be seen in
Section 3.2.
Here an analytical expression is derived for the reversible entrainment ratio of an ideal
gas process. The specific heat capacities of the gases at constant pressure are assumed to be
equal and constant. The equation of state for an ideal gas is:

P v = RT (9)
where P is the total pressure, v the specific volume, R the ideal gas constant on a unit mass
basis and T the total temperature. Equation (2) for ideal gases takes the following form for
a reversible ejector process:

cp TM + ER · cp TE = (1 + ER)cp TD (10)
Rearranging this equation, the discharge temperature may be written in terms of the inlet
conditions and the reversible entrainment ratio:
ER · TE + TM
TD = (11)
1 + ER
Rearranging Eq. (11) the following dimensionless equations are obtained:

TM 1 + ER
= (12)
TD 1 + ER · TTME
TM
TD T
+ ER
= E (13)
TE 1 + ER
Next, rearranging Eq. (7), the Second Law may be written as follows:
Ṡgen
sM − srev
D + ṁM
ER = (14)
(srev
D − sE )

Here we seek an expression linking the change in entropy to the state properties T and P .
The following expression describes a differential change in entropy of an ideal gas:
dT v
ds = cp − dP (15)
T T
dT dP
ds = cp −R (16)
T P

9
Now the entrainment ratio may be written as follows:
Ṡgen RM M
c dT − D R dP
R
ṁM
+ D p T P
ER = RD dT
R D dP (17)
c
E p T
− ERP

For constant specific heats Eq. (17) results in the following relation describing the reversible
entrainment ratio:
Ṡgen
ṁM
+ cp ln( TTM D
) − Rln( PPMD
)
ER = (18)
cp ln( TTDE ) − Rln( PPDE )
Dividing the right hand side of Eq. (18) by the ideal gas constant, R, renders all terms
dimensionless. Note also the definition, purely for convenience, of the dimensionless entropy
generation rate.

γ
s̄gen + γ−1
ln( TTMD
) − ln( PPMD
)
ER = γ (19)
γ−1
ln( TTDE ) − ln( PPDE )
Ṡgen
s̄gen = (20)
ṁM R
Eq. (12), (13) and Eq. (19) allow the entrainment ratio to be expressed as a function of
the following dimensionless parameters:

ER = fn(TE /TM , PE /PM , PD /PM , γ, s̄gen ) (21)

The reversible entrainment ratio, RER, can now be defined in terms of the inlet fluid
states and the discharge pressure using Eq. (19).
γ
γ−1
ln( TTM
D
) − ln( PPM
D
)
RER = γ (22)
γ−1
ln( TTDE ) − ln( PPDE )

3.2. Steam-steam Reversible Entrainment Ratios

In this section, the reversible entrainment ratio is calculated for steam-steam ejectors. As
a first step, let us ask how well steam-steam ejectors can be modelled using ideal gas ap-
proximations. We note that steam-steam ejectors are typically supplied with supersaturated
motive steam. This is done to avoid condensation within the ejector and a subsequent loss
of mass via a condensate stream.
Steam behaves like an ideal gas at conditions of high temperature or low pressure. Su-
persaturated steam is very well approximated by ideal gas assumptions at pressures below
approximately 10 bar. Above this pressure, ideal gas assumptions are only valid at higher
levels of superheat. In Fig. 8, the saturation curve of pure water is plotted on axes of specific
enthalpy versus specific entropy. The ratio of specific heats, γ, is calculated at a variety of
points within the superheated region and plotted to demonstrate that γ is approximately
constant in the low pressure superheated zone.
[Figure 8 about here.]

10
 The entrained pressure and discharge pressure for steam ejectors are typically close to or
below atmospheric pressure. Therefore, if the motive steam pressure is below about 10 bar,
Eq. (12), (13) and (21) are approximately applicable in describing the inlet and outlet states
of typical steam-steam ejectors.
The exact calculation of RER for steam is relatively straightforward. In each case the
equations solved are Eq. (6) & (7). (If the state lies under the saturation curve, T and
P would be insufficient to determine h and s. Instead, T and x, the dryness fraction, or
P and x would be required.) To graphically explain trends in the reversible entrainment
ratio of a steam-steam ejector, Fig. 9 is provided. Figure 9 shows the states of the motive,
discharge and entrained fluids on a specific enthalpy versus specific entropy diagram. The
state of the discharge fluid is given by the intersection of the isobar at the discharge pressure
and a line joining the motive and entrained states (again see Sect. 2 for derivation). The
entrainment ratio is represented by the length of the line from the motive to the discharge
state, |MD|, divided by the length of the line from the discharge to the entrained inlet state,
|DE|, Eq. (23).

|MD|
RER = (23)
|DE|
[Figure 9 about here.]

For a fixed state of the entrained fluid, it may be shown that a motive saturated steam
pressure exists that maximizes RER. In Fig. 10, RER is plotted as a function of the motive
steam pressure, choosing inlet fluid dryness fractions of 100% for illustrative purposes. RER
increases logarithmically until a motive steam pressure of approximately 15 bar is reached.
Beyond a motive saturated steam pressure of 15 bar, the rate of increase in RER reduces
until an optimal motive steam pressure is reached. This optimum pressure corresponds to the
point at which the tangent to the saturation curve is parallel to the slope of the isobar at the
discharge state (see Fig. 9). This conclusion is reached by considering the point at which the
distance |MD| divided by |DE| is maximised subject to point D lying at the intersection of
the isobar at the discharge pressure and a line joining states M and E. The optimal pressure
of the saturated motive fluid therefore depends upon the discharge pressure. As the discharge
pressure changes, the slope of the isobar changes and thus the point at which the tangent to
the saturation curve is tangential to the isobar at the discharge pressure also changes.

[Figure 10 about here.]

[Figure 11 about here.]

In Fig. 11, the optimal motive steam pressure is plotted for a fixed entrained fluid state and
discharge pressure. The results are primarily of academic interest given that the optimal
motive steam pressures are well beyond the typical pressures used in practical applications.
In Fig. 12 RER is plotted as a function of the discharge steam pressure. It is seen that
RER drops drastically with an increasing compression ratio. Often, the motive steam may be
superheated and therefore it is interesting to consider the effect of the degree of superheating
on RER. Figure 13 highlights the improvement in RER due to superheating at constant
pressure of the motive steam.

11
 [Figure 12 about here.]

[Figure 13 about here.]

In contrast to Fig. 10 and Fig. 12, the rate of change in RER is small with changes
in the inlet motive steam temperature. Again, this trend may be visualised making use of
Fig. 9. As the motive steam is superheated following an isobaric process, both its specific
enthalpy and its specific entropy increase, although its specific enthalpy increases at a greater
rate. This results in an increasing reversible entrainment ratio. The apparent linearity of
Fig. 13 may be explained by the fact that the isobars in the superheated region of Fig. 9 are
each approximately linear. Thus, the increase in reversible entrainment ratio, according to
Eq. (23), will be linear.
Finally, we consider how RER can be maximized when there is an upper limit on the
motive steam temperature. To do this we may plot RER at constant temperature against
motive steam specific entropy, as is done in Fig. 14.

[Figure 14 about here.]

The line plotted in Fig. 14 is a line of constant temperature. The reversible entrainment
ratio reaches a maximum when x=1, i.e. when the steam is in a saturated condition. Consider
the entrainment of saturated steam at state E by saturated motive steam at state M (see
Fig. 15). If the condition of the motive steam is brought below the saturation curve, the
ratio of the line |MD| to |DE| decreases. Thus, because the isobars on a h − s diagram of
steam converge as the state moves from that of saturated steam to that of a wet mixture,
RER using saturated motive steam will always be superior to that using a wet mixture at
the same temperature. In the superheated region of the h − s diagram, isobars and isotherms
do not coincide. If state M moves along an isotherm into the superheated region, RER
is severely compromised, as the isotherm corresponding to state M converges rapidly with
and meets the isobar corresponding to the discharge state D. In real steam-steam ejectors,
slightly superheated motive steam may preferentially be used to saturated steam in order to
discourage the formation of droplets in the discharge, which may need to be removed using
a mist eliminator.

[Figure 15 about here.]

As a final caveat, maximizing RER does not in any way maximize the exergetic efficiency
of a real device. The RER represents a perfect and reversible process that is 100% efficient
from a Second Law perspective. Rather, the purpose of this section was to comprehend the
upper thermodynamic limit on the entrainment ratio.

3.3. Steam-Moist-Gas Reversible Entrainment Ratios

Multipressure humidification-dehumidification (HDH) desalination cycles, which previ-
ously have been proposed by the current authors, make use of ejectors to pressurize a moist
carrier gas [13–15]. It is therefore of interest to consider the effect upon the reversible
entrainment ratio, RER, of using carrier gases of different types. In the case of a steam-
moist-carrier-gas ejector, the entrainment ratio is defined as follows:

12
 (1 + ωE )ṁcg
ER = (24)
ṁM
where ωE is the humidity ratio of the entrained fluid. In the case of humidification-dehumidification
desalination, the quantity of water vapor entrained is of importance, rather than the total
amount of fluid entrained. For this reason, it is useful to define the vapor entrainment ratio.
Of course, the vapor entrainment ratio is equivalent on a mass and molar basis as long as
the motive fluid is steam.
ωE ṁcg ωE
ERvap = = ER (25)
ṁM 1 + ωE
In this section the entrained and discharge fluids are treated as ideal mixtures of ideal
gases, within which the enthalpy and entropy of the carrier gas and steam are calculated
at their partial pressures. The ratio of the number of moles of carrier gas and water vapor
present in a moist carrier gas mixture along with the relative humidity are given as follows:

pvap Ṅst φpsat

ξ= = = (26)
P − pvap Ṅcg P − φpsat

When comparing the reversible entrainment ratio of different moist carrier gases, we do so
at equal entrained fluid temperature, pressure and relative humidity. Therefore, it is sensible
to consider the governing equations on a molar basis, since ξ is constant for all carrier gases
under these conditions.
First Law of Thermodynamics (reversible process):

Ṅst,M h̃st,M + Ṅcg h̃cg,E + Ṅst,E h̃st,E = Ṅst,D h̃rev rev

st,D + Ṅcg h̃cg,D (27)

where Ṅ is a molar flow rate and h̃ is a molar enthalpy. According to a mole balance on
steam:

Ṅst,D = Ṅst,M + Ṅst,E (28)

RERvap ERrev
vap
h̃st,M + h̃cg,E + RERvap h̃st,E = (1 + RERvap )h̃rev
st,D + h̃rev
cg,D (29)
ξE ξE

Second law of Thermodynamics (reversible process):

Ṅst,M s̃st,M + Ṅcg s̃cg,E + Ṅst,E s̃st,E = Ṅst,D s̃rev rev

st,D + Ṅcg s̃cg,D (30)
RERvap
s̃st,M + s̃cg,E + RERvap s̃st,E = (31)
ξE
RERvap rev
(1 + RERvap )s̃rev st,D + s̃cg,D (32)
ξE

13
Equations of State (reversible process):

h̃rev rev rev

st,D = fn(pst,D , s̃st,D ) (33)
h̃rev
cg,D = fn(prev rev
cg,D , s̃cg,D ) (34)

s̃ is the molar enthalpy. The partial pressures of the steam, pst,D , and carrier gas, pcg,D , both
sum to the total discharge pressure, PD . RER and RERvap are now plotted in Fig. 16 and
Fig. 17 as a function of motive steam pressure, setting xM and φE to unity for illustrative
purposes.2

[Figure 16 about here.]

[Figure 17 about here.]

The results of Fig. 16 collapse onto a single curve in Fig. 17. Surprisingly, RERvap appears
to be almost independent of carrier gas. Finally, in Fig. 18 the reversible vapor entrainment
ratio is plotted versus the entrained fluid inlet temperature, maintaining a relative humidity
of unity.

[Figure 18 about here.]

The increase in RERvap with temperature of the entrained moist carrier gas is explained
by the decrease in value of the specific heat of the entrained fluid per mole of vapor.

4. Development of Ejector Performance Metrics

When one seeks to evaluate the thermodynamic ideality of a process, performance metrics
are often chosen that compare useful work done in a real process to that in a defined reference
process. However, the identification of work done within an ejector is not a trivial matter. In
this section, these dilemmas are discussed and suggestions are made as to how performance
metrics can be defined.
The crux of difficulties in evaluating work done within an ejector is that one cannot easily
distinguish between useful flow work, heat transfer and dissipation. A similar difficulty is very
clearly explained by Casey and Fesich [16] in relation to diabatic turbochargers, where heat
transfer between the compressor and the turbine is significant. Maintaining the nomenclature
of Casey and Fesich, the Gibbs equation may be integrated along the process path of a single
fluid stream:
Z 2 Z 2 Z 2
h2 − h1 = v dp + (T ds)q + (T ds)irrev = y12 + q12 + j12 (35)
1 1 1
where y12 is the useful flow work done in the process, q12 is the reversible heat transfer and
j12 is the dissipation, or in other words, the lost work. Whilst the change in enthalpy of the

2
Note: Although not graphed here, the variation of reversible entrainment ratio with compression ratio for
helium, air and carbon dioxide carrier gases is qualitatively similar to that obtained for steam-steam ejectors.

14
fluid is fixed by the end states, the useful flow work, the heat transferred and the dissipation
are all path dependent. Only in the special case of an adiabatic and reversible process can
the useful flow work be equated with the change in enthalpy of the fluid:
Z 2
h2 − h1 = v dp (36)
1
In reality, the motive and entrained streams in an ejector remain neither adiabatic nor distinct
throughout the process. In real ejector processes, both heat transfer and dissipation are
experienced by the motive and entrained streams during mixing. Consequently, the useful
flow work can only be evaluated if the process path is known. Unfortunately, ejector mixing
processes are characterized by a high degree of macroscopic disorder, involving heat transfer
with finite temperature differences, forces that are not fully restrained and irreversible mixing.
Intermediate states cannot accurately be represented on a process path of one intensive
quantity against another. Consequently, there is no process path or equation in the form of
Eq. (35) that can be applied to actual ejector processes.
Despite the difficulty of identifying useful work done in an ejector, there remain two broad
approaches to characterising performance:

1. Comparisons may be made between terms representing useful work done in real and
reference processes, provided that a clear rationale is given as to why such a metric
compares different systems at different operating points. (Some performance metrics
may only evaluate the performance of a part of the entire ejector process, as shall be
seen with the turbine-compressor or compression efficiency.)
2. An exergy based approach may be taken. A performance metric may be defined once
the useful exergetic output, the exergetic input, and a reference state for the process
have been decided upon.

Next, four efficiencies based on the comparison of real and reversible processes are presented,
followed by an exergetic efficiency.

4.1. Reversible Entrainment Ratio Efficiency

The definition of reversible entrainment ratio efficiency was provided in work by Elrod
[1]. Al-Khalidy et al. [10] referenced the work of Elrod and also employed this efficiency.
The reversible entrainment ratio efficiency ηRER quantifies how close the actual entrainment
ratio of an ejector is to the theoretical maximum, for fixed conditions of operation. Its value
is therefore always between zero and unity. RER is calculated based upon the reversible
process described in Section 2.2.2.
ER
ηRER = (37)
RER
For a process involving fluids with identical ratios of specific heats and behaving as ideal
gases, the RER may be derived based upon the ideal gas analysis conducted in Section 3.1.1.

15
The ratio of the real to reversible entrainment ratio may be written as follows:
  
γ TM PM γ TM PM
ER  s̄gen + γ − 1 ln( TD ) − ln( PD )   γ − 1 ln( TD ) − ln( PD ) 
=   (38)
RER  γ TM PM  γ TM PM 
ln( ) − ln( ) ln( ) − ln( )
γ − 1 TD PD γ − 1 TD PD
One interesting case of ejector operation to consider is where the inlet temperatures of
both ideal gases are equal. According to the First Law, the discharge temperature must
equal the inlet temperatures in both real and reversible cases. Equation (22) then simplifies
to a ratio of the natural logarithm of the expansion ratio PM /PD to the ratio of the natural
logarithm of the compression ratio PD /PE .
PM
) ln(
PD
RER = (39)
PD
ln( )
PE
and the ratio of ER to RER simplifies to:
PM
ln( ) − s̄gen
ER PD
= (40)
RER PM
ln( )
PD
Finally, as a general comment, the principles of operation of steady flow ejectors impose
further limits upon ejector performance that are not captured by a control volume analysis
whereby the control volume surrounds the ejector. These internal limits depend both upon
ejector geometry and operational conditions. Some of the most enlightening work in this
regard has been conducted by Chou et al. [3] and Arbel et al. [4]. The same is true for the
reversible discharge pressure efficiency, to be presented in the following section.

4.2. Reversible Discharge Pressure Efficiency

The reversible discharge pressure efficiency quantifies how close the actual rise in pressure
of the entrained fluid is to the theoretical maximum pressure rise. As with the reversible
entrainment ratio efficiency, it is bounded between zero and unity. The theoretical rise in
pressure is calculated based upon Section 2.2.1. This efficiency is introduced and discussed
in detail by Arbel et al [4].
PD − PE
η= (41)
PDrev − PE

4.3. Turbine-Compressor Efficiency

The turbine-compressor efficiency [9, 17–19] (also known as the turbomachinery analogue
efficiency), allows the comparison of an ejector with a mechanically coupled turbine and
compressor. The efficiency is defined as the ratio of the entrainment ratio in a real ejector to
the entrainment ratio achieved in a turbo-charger exhausting to the same discharge pressure,

16
PD , as the ejector, and within which the turbine and compressor are adiabatic and isentropic
(Fig. 19).
ER
ηTER = (42)
TER
[Figure 19 about here.]

Here, TER stands for Turbo Entrainment Ratio or Turbine-Compressor Entrainment Ratio.
There is an important difference between the reversible entrainment ratio RER and the turbo
entrainment ratio TER. The RER is calculated based on a process whereby the discharge is
a single stream homogeneous in total temperature, total pressure and chemical composition.
The TER is based upon a process whereby the inlet streams remain distinct at the discharge,
i.e. the streams are not required to be in thermal or chemical equilibrium. Consequently,
the value of TER will always be less than or equal to RER. A corollary of this is that
ηTER is not bounded by an upper limit of 100%, as a real ejector could be imagined that
achieves an entrainment ratio greater than that of a turbo, through the exploitation of the
thermal or chemical disequilibrium of the fluids. Of course, a real steady-flow ejector could
not in practice achieve a turbine-compressor entrainment efficiency beyond 100%. There
is no obvious means by which a disequilibrium in chemical potential between the streams
may be exploited. The question as to how thermal disequilibria may be exploited is more
complex. As shall be seen in Sect. 5, for ejectors designed to date, the difference in the value
of ηTER and ηRER is small. Finally, ηTER may also be computed for a system comprising an
independent turbine and compressor with the role of expanding one fluid and compressing a
second, if the isentropic (also known as adiabatic) efficiencies of the turbine and compressor
are known. Bulusu et al. [9] demonstrates the following equality:
ER
ηTER = = ηturb,s ηcomp,s (43)
TER
where ηturb,s and ηcomp,s are the isentropic efficiencies of the turbine and compressor. Con-
sequently, it may be said that an ejector with a turbine-compressor efficiency of 35% will
provide the same performance as a turbine and compressor with isentropic efficiencies of 50%
and 70% respectively, for example.

4.4. Compression Efficiency

Al-Najem et al. [5] presented the following measure of ejector performance, which here is
denoted the compression efficiency:

(ṁM + ṁE ) (hD − hE )

ηc = (44)
ṁM (hM − hM,NE )
where hM,NE represents the static enthalpy at the exit of the motive fluid nozzle. Al-Najem
et al. mention that this efficiency compares the compression work done in the ejector to the
kinetic energy available from the motive stream. In a steady flow ejector, the nozzle through
which the motive fluid is accelerated is adiabatic. The quantity hM − hM,N E represents the
kinetic energy gained by the motive flow from the nozzle inlet to the exit. In an adiabatic
mechanical compressor, the compression work done is given by ṁE (hD −hE ). Based upon this

17
assertion, it seems likely that the quantity hD − hE was selected to represent the compression
work done in an ejector. However, in an ejector, heat transfer occurs between the motive
and entrained streams during mixing. Therefore, the change in enthalpy hD − hE is due to
a combination of heating/cooling and work. Indeed, the useful compression work done in an
ejector requires the thermodynamic states describing the process path of the entrained fluid
to be known. In conclusion, based upon the review of literature conducted, the rationale
behind and the utility of the compression efficiency is unclear.

4.5. Exergetic Efficiency

Exergetic analysis is a means of evaluating ejector performance from a Second Law point of
view. The premise of exergetic efficiency is to compare the useful exergetic output of a system
or component to the exergetic input. First, for an exergetic efficiency to be meaningful, an
environmental dead state must be defined. Let us consider an ejector operating within
an environment at T0 and P0 , as in Fig.2. Q̇0 represents the rate of heat input from the
surroundings. For a steady-state process, the First and Second Laws take the following form:

(hM − hD,M ) + Q̇0 = ER (hD,E − hE ) (45)

Q̇0
(sM − sD,M ) + + Ṡgen = ER (sD,E − sE ) (46)
T0
where states D,M and D,E represent the state of the motive and entrained fluids in the
discharge stream, calculated at their respective partial pressures. If the motive and entrained
fluids do not differ in chemical composition, these states will be identical. Multiplying the
Second Law by T0 and subtracting the resulting equation from the First Law, the ejector
process may be described in terms of flow exergy, x, a state variable.

(hM − hD ) − T0 (sM − sD ) = (hD − hE ) − T0 (sD − sE ) + T0 Ṡgen (47)

−∆ẋM = ER∆ẋE + T0 Ṡgen (48)
∆ẋE = (hD − hE ) − T0 (sD − sE ) (49)
∆ẋM = (hD − hM ) − T0 (sD − sM ) (50)

The next step is to identify the exergetic output from and exergetic input to the system.
This choice is somewhat subjective and depends on how the purpose of the ejector is defined.
It may be argued that the role of an ejector is to transfer exergy from the motive fluid to the
entrained fluid. This is somewhat analagous to a compressor or an extraction fan, where the
device’s role is to transfer exergy to a fluid. The value (or output) of the ejector process, in
exergetic terms, may be measured by the change in exergy of the entrained fluid. The price
paid to implement the process, again in exergetic terms, is the change in exergy of the motive
fluid. Hence, the following formulation for exergetic efficiency, identical to that employed by
Al-Najem et al. [5], is obtained:
ER∆ẋE
ηX = (51)
−∆ẋM

18
Considering Eq. (47) it is clear that ηX also quantifies the exergetic losses per unit of exergetic
input to the ejector process, Eq. (52).

T0 Ṡgen
1 − ηX = (52)
ṁM [(hD − hM ) − T0 (sD − sM )]

4.5.1. Comparison of the Exergetic and Reversible Entrainment Ejector Efficiencies

An analytical expression for the exergetic efficiency may be obtained when the inlet fluids
are ideal gases of identical and constant specific heats. If the further restriction of having
inlet fluids at the same temperature is imposed, the discharge enthalpy becomes independent
of the entrainment ratio, Eq. (11). The discharge temperature equals the inlet temperature,
and thus outlet and inlet specific enthalpies are also equal, Eq. (10). Consequently, the
exergetic ejector efficiency takes the following form.

(hD − T0 sD ) − (hE − T0 sE )
ηX = ER (53)
(hM − T0 sM ) − (hD − T0 sD )
−T0 sD + T0 sE
= ER (54)
−T0 sM + TsD
sD − sE
= ER (55)
sM − sD
PD
ln
PE
= ER (56)
PM
ln
PD
From Eq. (18), we an obtain the following expression for ER:

PM Ṡgen
ln( )−
PD RṁM
ER = (57)
PD
ln( )
PE
Upon substituting Eq. (57) into Eq. (56), the exergetic ejector efficiency is found to be equal
to the reversible entrainment ratio for an ideal gas ejector with identical specific heats and
identical inlet temperatures:

PM Ṡgen
ln( )−
PD RṁM
ηX = = ηRER (58)
PM
ln( )
PD

19
5. Comparison of Ejector Performance Metrics using Experimental Data
Now, having rigourously defined measures of ejector performance we may put these mea-
sures of efficiency to use in order to answer the following questions:

1. How do the numerical values of these efficiencies compare to one another?

2. What are the values of efficiency achieved by devices documented in literature?

To answer the first question, we can simply compare a sample of fictitious ejector perfor-
mance data. For example, we can compare the disparity in efficiency definitions over a range
of compression ratios, PD /PE . Let us consider a single case involving the compression of
saturated water vapor using high pressure saturated steam. Let us choose PM =10 bar and
PE =50 kPa. Let us fix the performance of the real ejector by fixing ER (Note: we could
also theoretically fix the real outlet state or the dimensionless entropy generation s̄gen ). Now,
in Fig. 20 the range of efficiencies are plotted versus the compression ratio, PD /PE . The
compression efficiency is omitted due to its lack of rigour and its requirement for knowledge
of a thermodynamic state internal to the ejector control volume.

[Figure 20 about here.]

To answer our second question, we must extract data from literature. One drawback of
using such data is that the performance of ejectors analysed in literature is not necessarily
representative of the state of the art ejectors employed in industry. Furthermore laboratory
tests may not have been performed at the ejectors’ design points of operation. With this
caveat in mind, the efficiencies of an air-air and a steam-steam ejector are analysed, acknowl-
edging that the conclusions drawn may not be generalised to cover all ejectors. Amenable
data for air-moist-carrier-gas ejectors was not available. Data for the air-air ejector was ex-
tracted from work by Bulusu et al. [9], Fig. 21 and Table 1 in Appendix 1. Data for the
steam-steam ejector is from work by Eames et al. [20], Fig. 22 and Table 2 in Appendix 2.
One should be aware that the degree of superheat of the motive steam is not obvious from
the reported work and motive steam has been assumed to be saturated at the entrance to
the ejector.

[Figure 21 about here.]

Note how, according to Eq. (58), the exergetic and reversible entrainment ratios are identical
in value. The discrepancy in the turbine-compressor efficiency compared to the former two
definitions is due to the fact that entropy generated in reaching thermal equilibrium of the
fluids is unaccounted for.

[Figure 22 about here.]

Note that critical conditions of operation typically refer to the mode of ejector operation
where the discharge pressure is such that choking of the entrained fluid occurs within the
ejector. This choking is due to the formation of a hypothetical throat for the entrained
fluid by the motive fluid expanding from the nozzle. For given inlet fluid conditions, the
critical point of operation corresponds to the point where the entrained mass flow rate is

20
maximum. The occurance of choking within steady flow ejectors is dealt with in great detail
by Chou et al. [3]. In Fig. 22 as in Fig. 21, the exergetic and reversible entrainment ratio
efficiencies are similar in value although both demonstrate an increasing trend. The turbine-
compressor efficiency is also close, and as expected, is greatest in value. Meanwhile the
reversible discharge pressure efficiency differs significantly in value. This may be explained in
two ways. Firstly, the reversible processes employed for the reversible discharge pressure and
the reversible entrainment ratio are different. Secondly, the ratio of mass flows, employed
in the reversible entrainment ratio, and the ratio of pressure rises of the entrained fluid,
employed in the reversible discharge pressure clearly do not scale in the same manner.

6. Conclusions
In this paper, various definitions of ejector efficiency have been evaluated. When defining
or employing an ejector efficiency, it is crucial to describe the physical significance of the
particular efficiency along with its utility. The following conclusions have been reached:

1. A reversible entrainment ratio may be calculated for any steady-flow ejector process,
by employing the the First and Second Laws of thermodynamics, provided the ther-
modynamic states of the inlet fluids and the total pressure of the discharge fluid, all at
equilibrium conditions, are known for a real ejector process.
2. The reversible entrainment ratio of an ideal gas ejector may be formulated analytically
as a function of the inlet (motive-to-entrained) temperature and pressure ratios, the
compression (discharge-to-entrained) pressure ratio and the ratio of specific heats. Cal-
culation of the real entrainment ratio also requires knowledge of a further state variable
describing the discharge (or alternatively the dimensionless entropy generation).
3. For steam-steam ejector applications, the inlet and outlet ejector states are shown to
be accurately represented by ideal gas equations for pressures below 10 bar. Higher
pressure motive steam may also be modelled as an ideal gas, provided the level of
superheat is sufficient.
4. Where the motive fluid employed in a ejector is saturated steam, there exists an optimal
pressure that maximizes the reversible entrainment ratio.
5. Where the temperature of motive steam is limited, the reversible entrainment ratio is
maximized by employing saturated motive steam at a pressure equal to the saturation
pressure at the maximum allowable temperature.
6. The reversible vapor entrainment ratio (ratio of vapor entrained to motive vapor) for
steam-moist-carrier gas ejectors is insensitive to the choice of carrier gas, if the inlet
relative humidity is maintained constant.
7. The turbine-compressor entrainment ratio differs from the reversible entrainment ratio
only in that the discharge fluids are unmixed and not brought into thermal and chemical
equilibrium.
8. The reversible entrainment ratio RER is always greater than or equal to the turbine-
compressor entrainment ratio TER.
9. The reversible entrainment ratio efficiency ηRER is always lower than the turbine-
compressor efficiency ηTER .

21
 10. For general air-air and steam-steam ejectors, the exergetic efficiency ηX is very close in
numerical value to the reversible entrainment ratio efficiency ηRER .
11. The exergetic efficiency ηX for an ideal gas ejector with inlet fluids at the same tem-
perature is identical to the reversible entrainment ratio efficiency ηRER .
12. The reversible entrainment ratio efficiency ηRER is advantageous compared to the ex-
ergetic efficiency ηX as its value does not depend upon the choice of a reference tem-
perature.

Acknowledgments
The authors would like to thank the King Fahd University of Petroleum and Minerals for
funding the research reported in this paper through the Center for Clean Water and Clean
Energy at MIT and KFUPM. The first author would like to acknowledge support provided
by the International Fulbright Science & Technology Award, U.S. Department of State.

22
7. References
[1] J. Elrod, H.G., The theory of ejectors, Journal of Applied Mechanics 12 (Sept. 1945)
A170–A174.

[2] J. Keenan, E. Neumann, F. Lustwerk, Investigation of ejector design by analysis and

experiment, American Society of Mechanical Engineers – Transactions – Journal of
Applied Mechanics 17 (1950) 299 – 309.

[3] S. Chou, P. Yang, C. Yap, Maximum mass flow ratio due to secondary flow choking in
an ejector refrigeration system, International Journal of Refrigeration 24 (2001) 486 –
499.

[4] A. Arbel, A. Shklyar, D. Hershgal, M. Barak, M. Sokolov, Ejector irreversibility charac-

teristics, Transactions of the ASME. Journal of Fluids Engineering 125 (2003/01/) 121
– 9.

[5] N. Al-Najem, M. A. Darwish, F. A. Youssef, Thermovapor compression desalters: energy

and availability - analysis of single- and multi-effect systems, Desalination 110 (1997)
223–238.

[6] O. A. Hamed, A. M. Zamamiri, S. Aly, N. Lior, Thermal performance and exergy

analysis of a thermal vapor compression desalination system, Energy Conversion and
Management 37 (1996) 379–387.

[7] A. O. B. Amer, Development and optimization of ME-TVC desalination system, De-

salination 249 (2009) 1315–1331.

[8] F. N. Alasfour, M. A. Darwish, A. O. B. Amer, Thermal analysis of ME-TVC+MEE

desalination systems, Desalination 174 (2005) 39–61.

[9] K. V. Bulusu, D. M. Gould, C. A. Garris Jr., Evaluation of efficiency in compressible flow

ejectors, in: 2008 ASME International Mechanical Engineering Congress and Exposition,
IMECE 2008, October 31, 2008 - November 6, volume 8, American Society of Mechanical
Engineers, Boston, MA, United States, 2008, pp. 531–554.

[10] N. Al-Khalidy, A. Zayonia, Design and experimental investigation of an ejector in an

air-conditioning and refrigeration system, in: Proceedings of the 1995 ASHRAE Annual
Meeting, June 24, 1995 - June 28, volume 101, Silesian Technical Univ, Gliwice, Poland,
ASHRAE, San Diego, CA, USA, 1995, pp. 383–391.

[11] H. El-Dessouky, H. Ettouney, I. Alatiqi, G. Al-Nuwaibit, Evaluation of steam jet ejectors,

Chemical Engineering and Processing 41 (2002) 551–561.

[12] F. Al-Juwayhel, H. El-Dessouky, H. Ettouney, Analysis of single-effect evaporator desali-

nation systems combined with vapor compression heat pumps, Desalination 114 (1997)
253 – 275.

23
[13] G. P. Narayan, R. K. McGovern, S. M. Zubair, J. H. Lienhard V, High-temperature-
steam-driven, varied-pressure, humidification-dehumidification system coupled with re-
verse osmosis for energy-efficient seawater desalination (2011). Manuscript under review.

[14] G. P. Narayan, M. H. Sharqawy, J. H. Lienhard V, S. M. Zubair, Water separation under

varied pressure, 2009. 2011/0079504 A1.

[15] G. P. Narayan, K. H. Mistry, J. H. Lienhard V, S. M. Zubair, High efficiency thermal

energy driven water purification system, 2011. USSN 13/028,170.

[16] M. V. Casey, T. M. Fesich, The efficiency of turbocharger compressors with diabatic

flows, Journal of Engineering for Gas Turbines and Power 132 (2010) 072302.

[17] I. Watanabe, Experimental investigations concerning pneumatic ejectors, with special

reference to the effect of dimensional parameters on performance characteristics, in: Jet
pumps and ejectors; Proceedings of the Symposium, London, England; United Kingdom;
1 Nov. 1972, 1972, pp. 97–120.

[18] K. R. Hedges, P. G. Hill, Compressible flow ejectors. II. Flow field measurements and
analysis, in: ASME-CSME Fluids Engineering Conference, volume 96, USA, 1974, pp.
282–8.

[19] K. Matsuo, K. Sasaguchi, Y. Kiyotoki, H. Mochizuki, Investigation of supersonic air

ejectors - II. Effects of throat-area-ratio on ejector performance, Bulletin of the JSME
25 (1982) 1898–1905.

[20] I. Eames, S. Aphornratana, H. Haider, Theoretical and experimental study of a small-

scale steam jet refrigerator, International Journal of Refrigeration 18 (1995) 378 – 386.

[21] K. V. Bulusu, Personal communication, 2011.

24
Appendix - Experimental Data Tables

25
 Table 1: Experimental air-air ejector data extracted from work by Bulusu et al., [9]
PM PE PD PDrev TM TE TD TDrev ER RER PIc s̄ ηRDP ηRER ηTME ηX
[kPa] [kPa] [kPa] [kPa] [K] [K] [K] [K] [-] [-] [-] [-] [%] [%] [%] [%]
769 27 105 620 293.7 293.7 292.3 293.2 0.068 1.46 3.92 1.90 13.18 4.69 7.51 4.69
677 30 104 538 293.7 293.7 292.5 293.3 0.079 1.50 3.48 1.77 14.64 5.28 8.18 5.28
601 36 104 492 293.7 293.1 292.6 293 0.076 1.65 2.90 1.67 14.92 4.62 6.85 4.62
538 42 103 453 293.7 293.1 292.7 293.1 0.072 1.84 2.45 1.58 14.87 3.93 5.60 3.93
485 48 103 417 293.7 293.1 292.8 293.1 0.071 2.03 2.14 1.49 14.94 3.47 4.78 3.47
437 63 103 387 294.3 292.6 293.4 292.9 0.067 2.94 1.63 1.41 12.34 2.27 2.96 2.27
394 68 103 354 294.3 292.6 293.5 292.9 0.066 3.28 1.51 1.32 12.09 2.00 2.54 2.00

26
356 72 103 320 294.3 292.6 293.6 292.9 0.072 3.48 1.43 1.22 12.43 2.08 2.58 2.08
322 75 102 294 294.3 292.6 293.7 292.9 0.068 3.75 1.36 1.12 12.36 1.82 2.22 1.82
292 78 102 268 294.3 292.6 293.7 292.9 0.069 3.94 1.31 1.03 12.58 1.74 2.08 1.74
264 79 102 237 294.3 292.0 293.7 292.5 0.098 3.76 1.29 0.92 14.44 2.60 3.05 2.59
238 83 102 218 294.3 292.0 293.8 292.4 0.093 4.14 1.23 0.83 14.05 2.26 2.60 2.25
215 87 102 201 294.3 292.0 293.8 292.4 0.085 4.66 1.17 0.73 13.29 1.83 2.06 1.82
194 90 102 184 294.3 292.0 293.9 292.4 0.076 5.20 1.13 0.63 12.66 1.46 1.61 1.45
174 93 102 167 294.3 292.0 294.0 292.3 0.072 5.96 1.09 0.53 11.86 1.21 1.32 1.21
155 95 102 150 294.3 292.0 294.0 292.3 0.065 6.45 1.07 0.42 11.67 1.00 1.06 1.00
129 97 102 127 294.3 291.5 294.0 291.9 0.060 5.72 1.04 0.24 14.05 1.04 1.08 1.03
 Table 2: Experimental steam-steam ejector data extracted from work by Eames et al., [20]
PM PE PD PDrev TD TDrev ER RER s̄ ηRDP ηRER ηTME ηX
[kPa] [kPa] [kPa] [kPa] [K] [K] [-] [-] [-] [%] [%] [%] [%]
199 0.87 3.40 27.5 344 304 0.59 3.1 0.59 9.5 19.1 21.3 19.9
232 0.87 3.70 34.0 349 306 0.54 3.0 0.54 8.5 18.2 20.4 19.1
270 0.87 4.40 44.7 355 310 0.47 2.6 0.47 8.1 17.9 20.3 18.8
313 0.87 5.10 62.1 362 313 0.39 2.4 0.39 6.9 16.1 18.6 17.2
362 0.87 5.40 89.3 370 314 0.31 2.4 0.31 5.1 13.0 15.0 14.0
199 1.04 3.60 35.6 349 305 0.50 3.3 0.50 7.4 15.0 16.5 15.8

27
232 1.04 4.10 48.4 356 308 0.42 3.0 0.42 6.5 13.8 15.3 14.7
270 1.04 4.60 63.8 362 310 0.36 2.8 0.36 5.7 12.7 14.2 13.6
313 1.04 5.10 89.0 369 312 0.29 2.7 0.29 4.6 10.8 12.2 11.7
362 1.04 5.70 123.6 376 315 0.23 2.5 0.23 3.8 9.1 10.4 10.0
199 1.23 3.80 47.7 355 305 0.40 3.6 0.40 5.5 11.0 12.0 11.9
232 1.23 4.19 63.0 361 307 0.34 3.4 0.34 4.8 10.0 11.0 10.9
270 1.23 4.70 85.0 367 310 0.28 3.1 0.28 4.1 8.9 9.9 9.8
313 1.23 5.30 105.6 372 312 0.25 2.9 0.25 3.9 8.6 9.7 9.5
362 1.23 6.00 154.3 380 315 0.18 2.7 0.18 3.1 6.7 7.6 7.5
 Discharge
Motive Fluid
NE
Fluid M D

E
Entrained Fluid
Figure 1: Schematic diagram of a steady flow ejector

28
 Motive Fluid
mM , PM , hM , sM Discharge
Fluid
mD , PD , hD , sD
Control Volume

Entrained Fluid
P0 ,T0 Q0
mE , PE , hE , sE
Figure 2: Control volume for an ejector process

29
 1
∝
h RER RER + 1
∝
RER + 1
Specific Enthalpy [J/kg]

PD = fn( sD , hD )

PM ,TM

RER
∝
RER + 1
PD ,TD
1
∝
RER + 1

PE ,TE

Specific Entropy [J/kg·K] s

Figure 3: Graphical representation of the reversible entrainment ratio

30
 h

PDrev
Specific Enthalpy [J/kg]

PD
PM ,TM
RER ER
∝ =
RER + 1 ER + 1
TDrev
TD
1 1
∝ =
RER + 1 ER + 1

PE ,TE

Specific Entropy [J/kg·K] s

Figure 4: Graphical representation of the reversible discharge pressure reference process

31
 h
Specific Enthalpy [J/kg]

PDrev = PD
PM ,TM
ER
∝ RER
TD ER + 1 ∝
RER + 1

TDrev
1
∝
RER + 1

PE ,TE 1
∝
ER + 1

Specific Entropy [J/kg·K] s

Figure 5: Graphical representation of the reversible entrainment ratio reference process

32
 E M

Wcomp Wturb
Compressor Turbine

WRHE
E′ M′

Reversible Heat Engine

Work flow
Heat flow D

Figure 6: Schematic diagram of a thermodynamically reversible ejector for fluids of identical chemical com-
position

33
 h PM

PD = PM' = PE'
Specific Enthalpy [J/kg]

M PE
E’
D

M’ E

Specific Entropy [J/kg·K] s

Figure 7: Process paths for a thermodynamically reversible ejector

34
 4000
P = 100 bar P = 10 bar P = 1 bar

P = 0.1 bar
Specific Enthalpy (kJ/kg)

3500
γ=1.28

γ= 1.29

γ= 1.30
3000
γ= 1.31
γ= 1.32

2500

Critical
Point

2000
4 5 6 7 8 9 10
Specific Entropy (kJ/kg·K)

Figure 8: Lines of constant ratio of specific heats, γ, for steam

35
 3000

100 bar 10 bar

2 bar
2800
Specific Enthalpy (kJ/kg)

M
1 bar
(hopt, sopt)
D E
2600

2400

parallel
2200
Critical
Point

2000
4 4.5 5 5.5 6 6.5 7 7.5
Specific Entropy (kJ/kg·K)

Figure 9: Specific enthalpy-entropy diagram for pure water

36
 16
Reversible Entrainment Ratio, RER

12

0
0 50 100 150 200
Motive Steam Pressure, PM [bar]

Figure 10: Reversible entrainment ratio versus motive steam pressure for a steam-steam ejector. (PD =15
kPa, PE =10 kPa, xM =1, xE =1)

37
 130
Motive Steam Pressure, PM,opt [bar]

125

120

115

110

105

100
0 20 40 60 80 100 120
Discharged Pressure, PD [kPa]

Figure 11: Optimal motive steam pressure of a steam-steam ejector

38
 Compression Ratio, Πc
0 2 4 6 8 10 12
50
Reversible Entrainment Ratio, RER

40

30

20

10

0
0 20 40 60 80 100 120
Discharged Pressure, PD [kPa]

Figure 12: Reversible entrainment ratio versus discharge pressure for a steam-steam ejector. (PM =10 bar,
PE =10 kPa, xM =1, xE =1)

39
 16

14
Reversible Entrainment Ratio, RER

12

10

0
0 25 50 75 100 125
Motive Steam Superheat, TM - TM,sat [°C]

Figure 13: Reversible entrainment ratio versus motive steam superheat for a steam-steam ejector. (PD =15
kPa, PM =10 bar, PE =10 kPa, xE =1, TM,sat = 180 ◦ C

40
 12

xM = 1
Reversible Entrainment Ratio, RER

10

Saturation Superheat
2 xM < 1 1 < xM

0
2 3 4 5 6 7 8
Motive Steam Specific Entropy, sM [kJ/kg·K]

Figure 14: Reversible entrainment ratio versus motive steam specific entropy for a steam-steam ejector, with
motive steam at constant temperature. (TM =179.9 ◦ C, Psat,M =10 bar, PD =15 kPa, PE =10 kPa, xE =1)

41
 3000

P = 2 bar

M
2800
Specific Enthalpy (kJ/kg)

T=453 K,
Psat=10 bar D
E
2600
T=393.4 K,
Psat=2 bar

2400
T=372.8 K,
Psat=1 bar

2200
Critical
Point

2000
4 5 6 7 8
Specific Entropy (kJ/kg·K)

Figure 15: Isotherms on a specific enthalpy-entropy diagram for pure water

42
 35
Helium Air Carbon Dioxide
30

25

20
RER

15

10

0
0 50 100 150 200
Motive Steam Pressure, PM [bar]

Figure 16: Reversible entrainment ratio versus motive steam pressure for steam-moist-carrier-gas ejectors.
(PD =75 kPa, PE =50 kPa, TE =50 ◦ C, xM =1, φE =1)

43
 4

3.5

2.5
RERvap

2
Carbon Dioxide

1.5 Air

Helium
1

0.5

0
0 50 100 150 200
Motive Steam Pressure, PM [bar]

Figure 17: Reversible vapor entrainment ratio versus motive steam pressure for steam-moist-carrier-gas
ejectors. (Πc =1.5, PE =50 kPa, TE =50 ◦ C, xM =1, φE =1)

44
 5

3
RERvap

2
Carbon Dioxide
Air

1 Helium

0
10 20 30 40 50 60 70 80
Entrained Carrier Gas Temperature, TE [°C]

Figure 18: Reversible vapor entrainment ratio versus entrained saturated carrier gas temperature for a steam-
moist-carrier-gas ejector. (PD =75 kPa, PE =50 kPa, Πi =20, xM =1, φE =1)

45
 h PM

PD = PM' = PE'
Specific Enthalpy [J/kg]

M PE
E’

M’
E

Specific Entropy [J/kg·K] s

Figure 19: Isentropic expansion and compression processes

46
 50

40
Efficiency [%]

30

20 X
TER
RER
10
RDP

0
1 1.1 1.2 1.3 1.4 1.5 1.6
Compression Ratio, Пc [-]

Figure 20: Comparison of efficiency definitions for a steam-steam ejector over a range of compression ratios

47
 16

14

12
Efficiency [%]

10
RDP
8 TER
X
6 RER

0
0 1 2 3 4 5
Compression Ratio, Пc [-]

Figure 21: Comparison of efficiency definitions using data from an air-air ejector, [9, 21]. Chamber throat
diameter, 0.88 inches; Diffuser exit diameter, 3.02 inches; Nozzle throat diameter, 0.5625 inches; Nozzle exit
area 0.735 inches; Plenum diameter for secondary flow, 3.25 inches.

48
 25

20
Efficiency [%]

TER
15
X
RER
RDP
10

0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Entrainment Ratio, ER [-]

Figure 22: Comparison of efficiency definitions using data from a steam-steam ejector operating at conditions
of critical back pressure [20]. Chamber throat diameter, 18 mm; Diffuser exit diameter, 40 mm; Nozzle throat
diameter, 2 mm; Nozzle exit area 8 mm; Plenum diameter for secondary flow, 24 mm.

49
Table 3: Reversible discharge pressure process
Fixed Output
PE = PErev PDrev
TE = TErev TDrev
rev
PM = PM
rev
TM = TM
ER = RER

50
Table 4: Reversible entrainment ratio process
Fixed Output
PE = PErev RER
TE = TErev TDrev
rev
PM = PM
rev
TM = TM
PD = PDrev

51